Both thermocouples and RTD's are in widespread use for sensing temperature and providing an electrical output representative of the temperature sensed. Thermocouples, by their nature, are point sensors because they thermoelectrically produce an electromotive force at a specific junction between two different materials. RTD's employ a wire sensing element which has a resistance which varies with temperature. Most RTD's are now designed to concentrate the electrical resistance to a small point or in the smallest possible volume, with miniaturization being a principle feature so that RTD's are, like thermocouples, essentially point sensors. Because of this point sensing feature, whenever an extended field is to be interrogated with the use of either thermocouples or RTD's, it has generally been necessary to distribute a multiplicity of differentially heated differential thermocouples or differential RTD's and measuring the heat loss at the particular point where the differential temperature is being sensed.
No matter how many point sensing thermocouples or RTD's are distributed in the field, they are unable to provide an accurate, true analog representation of the information to be determined from the field because they are still only sensing specific points. Determining the best points to interrogate, installing the individual sensor elements, and making the numerous required individual electrical connections to the point sensing elements in accordance with the generally accepted technology, are cumbersome and expensive steps.
There are many situations where it is desirable to sense level or flow rate over an extended field. This has been accomplished to a certain degree with thermocouples and RTD's by converting the point sensing reading to an average temperature of the field. However, the larger such a differential thermal field is, and the more varied the temperatures are across the field, the more point sensing elements are required to obtain a readout which is reasonably representative of the average condition of the field.
One situation where extended field interrogation is currently made with thermocouples or RTD's involves gauging of the fluid level, or the location of a phase change interface such as between liquid and gas, in a vessel such as a tank. This type of level gauging can currently be accomplished with thermocouples and RTD's by arranging a series of spaced sensor elements along the height of the tank, that is, at vertically separated points in the field being interrogated. In the case of RTD's, a series of heated RTD's and companion reference RTD's are employed along the height of the tank. As liquid reaches each RTD point sensor, the sensor reports that it is wet when it is cooled by the higher thermal dispersion rate of the liquid than is the case for the air or gas phase above the surface of the liquid. However, the operator is unable, with such point sensing structure, to determine whether the liquid level is just at that particular point or at any level between that point and just below the next higher RTD sensing point. Further filling of the tank will result in discrete reports from the sequentially higher RTD's, while lowering of the liquid level will cause successive discrete reports from the successively lower RTD's as they are uncovered from the liquid. For example, if ten sensing points are employed along the height of the tank, each with an individual heated RTD sensor and a reference RTD sensor, the gauging can only be performed at ten individual step points with unresolvable uncertainty of where the liquid level is between any two of those points. The only way to reduce such uncertainty when employing point sensors is to increase the number of sensing elements, at correspondingly increased expense, cumbersome wire connections and possibly reduced reliability.
Accurate liquid level sensing is of critical importance in any liquid storage vessels and particularly in reactor buildings of nuclear power plants, as well as in the reactor vessels themselves. This accurate liquid level sensing is important in avoiding nuclear power plant accidents which could be caused when the actual level of the liquid is either not properly known or is misinterpreted. In addition to lack of desired accuracy, liquid level changes are not immediately sensed when point sensors are used since there can be considerable change in liquid level prior to detection by the next sensing element which is either covered or uncovered. Thus, a developing problem or trend may not be immediately detected and the desirable mitigating action to suppress or correct the problem cannot be taken in as timely a manner as may be desirable or necessary.
Each of the vertical sequence of thermocouples or RTD's in such a liquid level gauging system requires its own separate electrical connections to the detection circuitry. The thereby required large number of joints or splices can result in undesirably low reliability, which could be especially dangerous in the environment of a nuclear power plant. As an example of this problem, there is in existence a point-sensing RTD system for water level sensing in a nuclear reactor building which has approximately fifty RTD sensors arrayed over a vertical height of about sixty feet.
RTD's are preferred for some purposes over thermocouples because they can be made more sensitive, being able to provide an output signal many times greater than is generally possible with thermocouples. This is primarily because RTD's operate with an external electrical power source which can provide as high a level of voltage or current as is desired. Thermocouples operate on the basis of a self-generated junction electromotive force (emf) which inherently has a very low output voltage level as well as other inaccuracies.
For sensing in some extended fields, such as the inside of a nuclear reactor vessel, access may be relatively difficult and may be best achieved by encasing a series of sensors in a long, slender, tubular probe. Such a probe can be readily inserted in an existing reactor vessel instrument guide tube. RTD's are desirable in such situations because of their high output and therefore high sensitivity but many prior art RTD's are not suitable for such packaging, being too bulky and having a ceramic or a glass insulator too brittle to allow them to be deformed as would be required for packaging in a long, slender, tubular probe.
On the other hand, thermocouples have been packaged inside a metal casing as small as 0.01 inch in diameter. A series of such encased thermocouples and the required electrical leads may be placed inside a tube and encased by drawing or swaging the tube down around the thermocouples and leads to produce a long slender probe suitable for gaining access in restricted regions inside a nuclear reactor vessel. This advantage for the thermocouples is balanced by at least one equivalent disadvantage. Thermocouples are relatively delicate and are easily subject to breakage during the manufacture of such probes or upon accidental impacting. Of course, as discussed above, a thermocouple being a point sensor, thermocouple probes necessarily have a step function output rather than a continuous output, so the accuracy of liquid level determination is limited. Additionally, the electrical output of thermocouples is so small that performance is grainy and resolution and accuracy are relatively poor. Because individual wire leads are required for each thermocouple, numerous wires must extend along the probe, thereby limiting how small the outside diameter of a tubular probe can be. Of course, the greater the number of thermocouples placed along the probe in any attempt to increase resolution the greater the number of leads. This large number of leads also seriously reduces the reliability of thermocouple-type probes. Thermocouple probes are relatively expensive to make, especially considering the number of leads, connections and electronic cooperating devices required for such probes
Another example of an extended field which has been interrogated by a multiplicity of RTD's or thermocouple sensors, is a large duct having a non-uniform flow profile, where it is sought to obtain an average reading of the flow velocity in the duct. Such non-uniform flow distributions exist, for example, in air ducts where diameters are large and fittings such as tees, elbows, transitions, bends, section changes, louvers, dampers and the like cause flow disturbances Non-uniform flow distributions also typically occur in the input air ducts and combustion output ducts of fossil fuel power plants. In such cases, a multiplicity of point sensing elements is placed at what are considered to be strategic locations across the gas flow path, but only a rough approximation of the flowable velocity can be obtained by the use of such discrete, point sensing locations. As stated previously, a large number of individual point sensors results in high costs due to the number of leads, connectors and mating electronic devices that are necessary to cooperate in interpreting the individual signals.